Images taken on digital cameras or the like are subjected to image retouching (correction) to revise a part having a defect and eliminate the defect. Such retouching is performed by, for example, changing image representation, such as color tone and sharpness.
There are the following two types of image retouching: a type in which a user or an apparatus specifies away, a parameter, or the like of image processing and retouching is accordingly performed (hereinafter, referred to as a standard type); and a type in which a feature amount, or the like is specified and retouching is performed toward the target (hereinafter, referred to as a transfer type). Besides the classification by the retouching type, there are also the following three types by the contents specified by the user.
The first type is a type in which the user directly specifies an operation (an operator) and parameters of image retouching (hereinafter, referred to as a way specification type). The second type is a type in which the user specifies a target in an image format or the like (hereinafter, referred to as a target specification type). The third type is a type in which the user specifies a retouched image that finally meets his/her expectation from a plurality of retouched images (hereinafter, referred to as a result specification type).
By combining the former two types and the latter three types, six types in total are conceivable. Because there is no method that can satisfy both the transfer type and the way specification type, there are actually five types.
Examples of the standard type include tone curve conversion, sharpness conversion, and edge enhancement. Because these ways are publicly known, detailed explanations thereof are omitted. In these ways, the parameters can be optionally changed, thereby providing high processing flexibility. To achieve a target, however, trial and error is required, and proficiency is required. Examples of the combination of the standard type and the way specification type include a technique for causing a user to designate colors of designated pixels and desired colors thereof and adjusting colors of the entire image based on these colors (refer to Patent Literature 1).
There has also been developed a technique for creating a color conversion list indicating a correspondence relation between pixel values before and after adjustment based on the adjustment result of image data and applying the color conversion list to another image, thereby correcting the image (refer to Patent Literature 2). There have also been developed a technique for facilitating creation of a color conversion list including user interaction (refer to Patent Literature 3) and a similar technique assuming video as an object to be processed (refer to Patent Literature 4).
Examples of the combination of the standard type and the target specification type include a technique for displaying effects in an icon form to a user and causing the user to select an effect, which is performed by an application called Instagram. Examples of the combination of the standard type and the result specification type include a technique for specifying a small number of typical processing and parameters in advance, displaying these and retouched images in pairs to a user, and causing the user to select a retouched image (refer to Patent Literature 5).
Examples of the combination of the transfer type and the target specification type include a technique for matching image characteristics (e.g., highlight density, shadow density, and average density) with those of a reference image to uniform the color tone or the like of finished photo prints of similar scenes (refer to Patent Literature 6). There has also been developed a technique for causing a user to select a target face image (area) and matching the color tone of a color area of a face image in an input image to that of the color area of the target face image to uniform the color tone of the face of the same subject in a digital photo printer (refer to Patent Literature 7). There has also been developed a technique for correcting image reproduction items (color reproduction, tone reproduction, and contour reproduction) of an input image such that the image reproduction items approach those of reference teacher data for intuitively performing desired image reproduction (refer to Patent Literature 8).
Examples of the combination of the transfer type and the result specification type include a technique for selecting a desired first image from results obtained by correction processing performed such that a second feature amount of a second image to be corrected approaches respective first feature amounts of a plurality of types of first images (refer to Patent Literature 9).
The combination of the standard type and the way specification type provides simple options to the user at the expense of the processing flexibility, which is an advantage of the standard type, thereby limiting the processing flexibility. Furthermore, the combination has difficulty in prediction in that how the overall impression of the image is changed by a result of application of color conversion in pixel units cannot be seen until the processing is actually performed. In the combination of the standard type and the target specification type, the actual effects depend on the input image. Thus, the combination also has difficulty in prediction and requires trial and error. The combination of the standard type and the result specification type has no difficulty in prediction because the user can select an image from the retouched images. The combination, however, provides options to the user at the expense of the processing flexibility, thereby limiting the processing flexibility.
The combination of the transfer type and the target specification type does not allow the user to optionally change the way and parameter, thereby extremely limiting the processing flexibility. The combination can alleviate the difficulty in prediction compared with the combinations described above but cannot eliminate it. The combination of the transfer type and the result specification type can eliminate the difficulty in prediction but cannot solve the problem of limiting the processing flexibility similarly to the combination of the transfer type and the target specification type.
As described above, the combinations of two types can eliminate the difficulty in prediction compared with the standard type alone but limit the processing flexibility. If the standard type alone is used to increase the processing flexibility, the difficulty in prediction occurs. If the difficulty in prediction does not arise, the intuitiveness and the usability are increased and a desired result can be obtained by small workload. Thus, it is desired to provide a system and a method that can achieve intuitive and simple use, reach a desired result by small workload, and increase the processing flexibility.